Developing a New Measure of the Useful Field of View for Use in Dynamic Real-World Scene Viewing

Abstract

In real-world contexts, such as driving, a person’s breadth of attention, or useful field of view (UFOV), can have life or death consequences, with a narrower UFOV associated with increased accident risk (e.g., Clay et al., 2005). However, existing measures of the UFOV have important limitations. Some cannot be used in dynamic viewing of real-world scenes, while others (e.g., peripheral detection tasks) do not control for retinal eccentricity or eccentricity-dependent contrast sensitivity. The current experiment aimed to develop a novel measure of the UFOV that overcomes these limitations. Our dependent measure was the detection of extrafoveal image blur in real-world scenes as a secondary task, while participants concurrently engaged in an attention-demanding primary task. The retinal eccentricity of the image blur was controlled through gaze-contingent presentation on occasional single fixations. Eccentricity-dependent contrast sensitivity was held constant in the following way. Blurred images contained a circular region of high resolution (3°, 6°, or 9° radius) centered on fixation, with a constant level of low-pass filtered imagery beyond that eccentricity. Each eccentricity was paired with a unique blur level such that, in a single-task blur detection task, blur detectability was held constant across eccentricities. Our first experiment, which disallowed eye movements, used a within-subjects design (n = 16) and occasional briefly flashed photographs of real-world scenes, half of which were blurred, while monitoring the viewer's eyes to ensure central fixation. To measure the effects of cognitive load on blur detection, participants concurrently did an auditory N-back task (with N = 0, 2, or 3). Results showed that as N-back level increased, blur detection significantly decreased, but did not interact with eccentricity—consistent with a general interference effect rather than tunnel vision (Crundall, Underwood & Chapman, 1999). Follow-up experiments will allow free viewing of scenes, and occasionally present blur for